4ba9b64b |
1 | /* |
2 | * Zlib (RFC1950 / RFC1951) compression for PuTTY. |
3 | * |
4 | * There will no doubt be criticism of my decision to reimplement |
5 | * Zlib compression from scratch instead of using the existing zlib |
6 | * code. People will cry `reinventing the wheel'; they'll claim |
7 | * that the `fundamental basis of OSS' is code reuse; they'll want |
8 | * to see a really good reason for me having chosen not to use the |
9 | * existing code. |
10 | * |
11 | * Well, here are my reasons. Firstly, I don't want to link the |
12 | * whole of zlib into the PuTTY binary; PuTTY is justifiably proud |
13 | * of its small size and I think zlib contains a lot of unnecessary |
14 | * baggage for the kind of compression that SSH requires. |
15 | * |
16 | * Secondly, I also don't like the alternative of using zlib.dll. |
17 | * Another thing PuTTY is justifiably proud of is its ease of |
18 | * installation, and the last thing I want to do is to start |
19 | * mandating DLLs. Not only that, but there are two _kinds_ of |
20 | * zlib.dll kicking around, one with C calling conventions on the |
21 | * exported functions and another with WINAPI conventions, and |
22 | * there would be a significant danger of getting the wrong one. |
23 | * |
24 | * Thirdly, there seems to be a difference of opinion on the IETF |
25 | * secsh mailing list about the correct way to round off a |
26 | * compressed packet and start the next. In particular, there's |
27 | * some talk of switching to a mechanism zlib isn't currently |
28 | * capable of supporting (see below for an explanation). Given that |
29 | * sort of uncertainty, I thought it might be better to have code |
30 | * that will support even the zlib-incompatible worst case. |
31 | * |
32 | * Fourthly, it's a _second implementation_. Second implementations |
33 | * are fundamentally a Good Thing in standardisation efforts. The |
34 | * difference of opinion mentioned above has arisen _precisely_ |
35 | * because there has been only one zlib implementation and |
36 | * everybody has used it. I don't intend that this should happen |
37 | * again. |
38 | */ |
39 | |
40 | #include <stdlib.h> |
41 | #include <assert.h> |
42 | |
52912ff3 |
43 | #ifdef ZLIB_STANDALONE |
44 | |
45 | /* |
46 | * This module also makes a handy zlib decoding tool for when |
47 | * you're picking apart Zip files or PDFs or PNGs. If you compile |
48 | * it with ZLIB_STANDALONE defined, it builds on its own and |
49 | * becomes a command-line utility. |
50 | * |
51 | * Therefore, here I provide a self-contained implementation of the |
52 | * macros required from the rest of the PuTTY sources. |
53 | */ |
54 | #define snew(type) ( (type *) malloc(sizeof(type)) ) |
55 | #define snewn(n, type) ( (type *) malloc((n) * sizeof(type)) ) |
56 | #define sresize(x, n, type) ( (type *) realloc((x), (n) * sizeof(type)) ) |
57 | #define sfree(x) ( free((x)) ) |
58 | |
59 | #else |
4ba9b64b |
60 | #include "ssh.h" |
52912ff3 |
61 | #endif |
4ba9b64b |
62 | |
dc3c8261 |
63 | #ifndef FALSE |
64 | #define FALSE 0 |
65 | #define TRUE (!FALSE) |
66 | #endif |
67 | |
4ba9b64b |
68 | /* ---------------------------------------------------------------------- |
69 | * Basic LZ77 code. This bit is designed modularly, so it could be |
70 | * ripped out and used in a different LZ77 compressor. Go to it, |
71 | * and good luck :-) |
72 | */ |
73 | |
74 | struct LZ77InternalContext; |
75 | struct LZ77Context { |
76 | struct LZ77InternalContext *ictx; |
77 | void *userdata; |
32874aea |
78 | void (*literal) (struct LZ77Context * ctx, unsigned char c); |
79 | void (*match) (struct LZ77Context * ctx, int distance, int len); |
4ba9b64b |
80 | }; |
81 | |
82 | /* |
83 | * Initialise the private fields of an LZ77Context. It's up to the |
84 | * user to initialise the public fields. |
85 | */ |
86 | static int lz77_init(struct LZ77Context *ctx); |
87 | |
88 | /* |
89 | * Supply data to be compressed. Will update the private fields of |
90 | * the LZ77Context, and will call literal() and match() to output. |
6e9e9520 |
91 | * If `compress' is FALSE, it will never emit a match, but will |
92 | * instead call literal() for everything. |
4ba9b64b |
93 | */ |
94 | static void lz77_compress(struct LZ77Context *ctx, |
32874aea |
95 | unsigned char *data, int len, int compress); |
4ba9b64b |
96 | |
97 | /* |
98 | * Modifiable parameters. |
99 | */ |
100 | #define WINSIZE 32768 /* window size. Must be power of 2! */ |
101 | #define HASHMAX 2039 /* one more than max hash value */ |
102 | #define MAXMATCH 32 /* how many matches we track */ |
103 | #define HASHCHARS 3 /* how many chars make a hash */ |
104 | |
105 | /* |
106 | * This compressor takes a less slapdash approach than the |
107 | * gzip/zlib one. Rather than allowing our hash chains to fall into |
108 | * disuse near the far end, we keep them doubly linked so we can |
109 | * _find_ the far end, and then every time we add a new byte to the |
110 | * window (thus rolling round by one and removing the previous |
111 | * byte), we can carefully remove the hash chain entry. |
112 | */ |
113 | |
114 | #define INVALID -1 /* invalid hash _and_ invalid offset */ |
115 | struct WindowEntry { |
d2371c81 |
116 | short next, prev; /* array indices within the window */ |
117 | short hashval; |
4ba9b64b |
118 | }; |
119 | |
120 | struct HashEntry { |
d2371c81 |
121 | short first; /* window index of first in chain */ |
4ba9b64b |
122 | }; |
123 | |
124 | struct Match { |
125 | int distance, len; |
126 | }; |
127 | |
128 | struct LZ77InternalContext { |
129 | struct WindowEntry win[WINSIZE]; |
130 | unsigned char data[WINSIZE]; |
131 | int winpos; |
132 | struct HashEntry hashtab[HASHMAX]; |
133 | unsigned char pending[HASHCHARS]; |
134 | int npending; |
135 | }; |
136 | |
32874aea |
137 | static int lz77_hash(unsigned char *data) |
138 | { |
139 | return (257 * data[0] + 263 * data[1] + 269 * data[2]) % HASHMAX; |
4ba9b64b |
140 | } |
141 | |
32874aea |
142 | static int lz77_init(struct LZ77Context *ctx) |
143 | { |
4ba9b64b |
144 | struct LZ77InternalContext *st; |
145 | int i; |
146 | |
3d88e64d |
147 | st = snew(struct LZ77InternalContext); |
4ba9b64b |
148 | if (!st) |
149 | return 0; |
150 | |
151 | ctx->ictx = st; |
152 | |
153 | for (i = 0; i < WINSIZE; i++) |
154 | st->win[i].next = st->win[i].prev = st->win[i].hashval = INVALID; |
155 | for (i = 0; i < HASHMAX; i++) |
156 | st->hashtab[i].first = INVALID; |
157 | st->winpos = 0; |
158 | |
159 | st->npending = 0; |
160 | |
161 | return 1; |
162 | } |
163 | |
164 | static void lz77_advance(struct LZ77InternalContext *st, |
32874aea |
165 | unsigned char c, int hash) |
166 | { |
4ba9b64b |
167 | int off; |
168 | |
169 | /* |
170 | * Remove the hash entry at winpos from the tail of its chain, |
171 | * or empty the chain if it's the only thing on the chain. |
172 | */ |
173 | if (st->win[st->winpos].prev != INVALID) { |
174 | st->win[st->win[st->winpos].prev].next = INVALID; |
175 | } else if (st->win[st->winpos].hashval != INVALID) { |
176 | st->hashtab[st->win[st->winpos].hashval].first = INVALID; |
177 | } |
178 | |
179 | /* |
180 | * Create a new entry at winpos and add it to the head of its |
181 | * hash chain. |
182 | */ |
183 | st->win[st->winpos].hashval = hash; |
184 | st->win[st->winpos].prev = INVALID; |
185 | off = st->win[st->winpos].next = st->hashtab[hash].first; |
186 | st->hashtab[hash].first = st->winpos; |
187 | if (off != INVALID) |
188 | st->win[off].prev = st->winpos; |
189 | st->data[st->winpos] = c; |
190 | |
191 | /* |
192 | * Advance the window pointer. |
193 | */ |
32874aea |
194 | st->winpos = (st->winpos + 1) & (WINSIZE - 1); |
4ba9b64b |
195 | } |
196 | |
197 | #define CHARAT(k) ( (k)<0 ? st->data[(st->winpos+k)&(WINSIZE-1)] : data[k] ) |
198 | |
199 | static void lz77_compress(struct LZ77Context *ctx, |
32874aea |
200 | unsigned char *data, int len, int compress) |
201 | { |
4ba9b64b |
202 | struct LZ77InternalContext *st = ctx->ictx; |
203 | int i, hash, distance, off, nmatch, matchlen, advance; |
204 | struct Match defermatch, matches[MAXMATCH]; |
205 | int deferchr; |
206 | |
207 | /* |
208 | * Add any pending characters from last time to the window. (We |
209 | * might not be able to.) |
210 | */ |
211 | for (i = 0; i < st->npending; i++) { |
212 | unsigned char foo[HASHCHARS]; |
213 | int j; |
214 | if (len + st->npending - i < HASHCHARS) { |
215 | /* Update the pending array. */ |
216 | for (j = i; j < st->npending; j++) |
32874aea |
217 | st->pending[j - i] = st->pending[j]; |
4ba9b64b |
218 | break; |
219 | } |
220 | for (j = 0; j < HASHCHARS; j++) |
32874aea |
221 | foo[j] = (i + j < st->npending ? st->pending[i + j] : |
4ba9b64b |
222 | data[i + j - st->npending]); |
223 | lz77_advance(st, foo[0], lz77_hash(foo)); |
224 | } |
225 | st->npending -= i; |
226 | |
8a9977e5 |
227 | defermatch.distance = 0; /* appease compiler */ |
4ba9b64b |
228 | defermatch.len = 0; |
2d466ffd |
229 | deferchr = '\0'; |
4ba9b64b |
230 | while (len > 0) { |
231 | |
32874aea |
232 | /* Don't even look for a match, if we're not compressing. */ |
233 | if (compress && len >= HASHCHARS) { |
234 | /* |
235 | * Hash the next few characters. |
236 | */ |
237 | hash = lz77_hash(data); |
238 | |
239 | /* |
240 | * Look the hash up in the corresponding hash chain and see |
241 | * what we can find. |
242 | */ |
243 | nmatch = 0; |
244 | for (off = st->hashtab[hash].first; |
245 | off != INVALID; off = st->win[off].next) { |
246 | /* distance = 1 if off == st->winpos-1 */ |
247 | /* distance = WINSIZE if off == st->winpos */ |
248 | distance = |
249 | WINSIZE - (off + WINSIZE - st->winpos) % WINSIZE; |
250 | for (i = 0; i < HASHCHARS; i++) |
251 | if (CHARAT(i) != CHARAT(i - distance)) |
252 | break; |
253 | if (i == HASHCHARS) { |
254 | matches[nmatch].distance = distance; |
255 | matches[nmatch].len = 3; |
256 | if (++nmatch >= MAXMATCH) |
257 | break; |
258 | } |
259 | } |
260 | } else { |
261 | nmatch = 0; |
4ba9b64b |
262 | hash = INVALID; |
263 | } |
264 | |
265 | if (nmatch > 0) { |
266 | /* |
267 | * We've now filled up matches[] with nmatch potential |
268 | * matches. Follow them down to find the longest. (We |
269 | * assume here that it's always worth favouring a |
270 | * longer match over a shorter one.) |
271 | */ |
272 | matchlen = HASHCHARS; |
273 | while (matchlen < len) { |
274 | int j; |
275 | for (i = j = 0; i < nmatch; i++) { |
276 | if (CHARAT(matchlen) == |
277 | CHARAT(matchlen - matches[i].distance)) { |
278 | matches[j++] = matches[i]; |
279 | } |
280 | } |
281 | if (j == 0) |
282 | break; |
283 | matchlen++; |
284 | nmatch = j; |
285 | } |
286 | |
287 | /* |
288 | * We've now got all the longest matches. We favour the |
289 | * shorter distances, which means we go with matches[0]. |
290 | * So see if we want to defer it or throw it away. |
291 | */ |
292 | matches[0].len = matchlen; |
293 | if (defermatch.len > 0) { |
294 | if (matches[0].len > defermatch.len + 1) { |
295 | /* We have a better match. Emit the deferred char, |
296 | * and defer this match. */ |
32874aea |
297 | ctx->literal(ctx, (unsigned char) deferchr); |
4ba9b64b |
298 | defermatch = matches[0]; |
299 | deferchr = data[0]; |
300 | advance = 1; |
301 | } else { |
302 | /* We don't have a better match. Do the deferred one. */ |
303 | ctx->match(ctx, defermatch.distance, defermatch.len); |
304 | advance = defermatch.len - 1; |
305 | defermatch.len = 0; |
306 | } |
307 | } else { |
308 | /* There was no deferred match. Defer this one. */ |
309 | defermatch = matches[0]; |
310 | deferchr = data[0]; |
311 | advance = 1; |
32874aea |
312 | } |
4ba9b64b |
313 | } else { |
314 | /* |
315 | * We found no matches. Emit the deferred match, if |
316 | * any; otherwise emit a literal. |
317 | */ |
318 | if (defermatch.len > 0) { |
319 | ctx->match(ctx, defermatch.distance, defermatch.len); |
320 | advance = defermatch.len - 1; |
321 | defermatch.len = 0; |
322 | } else { |
323 | ctx->literal(ctx, data[0]); |
324 | advance = 1; |
325 | } |
326 | } |
327 | |
328 | /* |
329 | * Now advance the position by `advance' characters, |
330 | * keeping the window and hash chains consistent. |
331 | */ |
332 | while (advance > 0) { |
333 | if (len >= HASHCHARS) { |
334 | lz77_advance(st, *data, lz77_hash(data)); |
335 | } else { |
336 | st->pending[st->npending++] = *data; |
337 | } |
338 | data++; |
339 | len--; |
340 | advance--; |
341 | } |
342 | } |
343 | } |
344 | |
345 | /* ---------------------------------------------------------------------- |
346 | * Zlib compression. We always use the static Huffman tree option. |
347 | * Mostly this is because it's hard to scan a block in advance to |
348 | * work out better trees; dynamic trees are great when you're |
349 | * compressing a large file under no significant time constraint, |
350 | * but when you're compressing little bits in real time, things get |
351 | * hairier. |
352 | * |
353 | * I suppose it's possible that I could compute Huffman trees based |
354 | * on the frequencies in the _previous_ block, as a sort of |
355 | * heuristic, but I'm not confident that the gain would balance out |
356 | * having to transmit the trees. |
357 | */ |
358 | |
4ba9b64b |
359 | struct Outbuf { |
360 | unsigned char *outbuf; |
361 | int outlen, outsize; |
362 | unsigned long outbits; |
363 | int noutbits; |
364 | int firstblock; |
6e9e9520 |
365 | int comp_disabled; |
4ba9b64b |
366 | }; |
367 | |
32874aea |
368 | static void outbits(struct Outbuf *out, unsigned long bits, int nbits) |
369 | { |
4ba9b64b |
370 | assert(out->noutbits + nbits <= 32); |
371 | out->outbits |= bits << out->noutbits; |
372 | out->noutbits += nbits; |
373 | while (out->noutbits >= 8) { |
32874aea |
374 | if (out->outlen >= out->outsize) { |
375 | out->outsize = out->outlen + 64; |
3d88e64d |
376 | out->outbuf = sresize(out->outbuf, out->outsize, unsigned char); |
32874aea |
377 | } |
378 | out->outbuf[out->outlen++] = (unsigned char) (out->outbits & 0xFF); |
379 | out->outbits >>= 8; |
380 | out->noutbits -= 8; |
4ba9b64b |
381 | } |
382 | } |
383 | |
384 | static const unsigned char mirrorbytes[256] = { |
385 | 0x00, 0x80, 0x40, 0xc0, 0x20, 0xa0, 0x60, 0xe0, |
386 | 0x10, 0x90, 0x50, 0xd0, 0x30, 0xb0, 0x70, 0xf0, |
387 | 0x08, 0x88, 0x48, 0xc8, 0x28, 0xa8, 0x68, 0xe8, |
388 | 0x18, 0x98, 0x58, 0xd8, 0x38, 0xb8, 0x78, 0xf8, |
389 | 0x04, 0x84, 0x44, 0xc4, 0x24, 0xa4, 0x64, 0xe4, |
390 | 0x14, 0x94, 0x54, 0xd4, 0x34, 0xb4, 0x74, 0xf4, |
391 | 0x0c, 0x8c, 0x4c, 0xcc, 0x2c, 0xac, 0x6c, 0xec, |
392 | 0x1c, 0x9c, 0x5c, 0xdc, 0x3c, 0xbc, 0x7c, 0xfc, |
393 | 0x02, 0x82, 0x42, 0xc2, 0x22, 0xa2, 0x62, 0xe2, |
394 | 0x12, 0x92, 0x52, 0xd2, 0x32, 0xb2, 0x72, 0xf2, |
395 | 0x0a, 0x8a, 0x4a, 0xca, 0x2a, 0xaa, 0x6a, 0xea, |
396 | 0x1a, 0x9a, 0x5a, 0xda, 0x3a, 0xba, 0x7a, 0xfa, |
397 | 0x06, 0x86, 0x46, 0xc6, 0x26, 0xa6, 0x66, 0xe6, |
398 | 0x16, 0x96, 0x56, 0xd6, 0x36, 0xb6, 0x76, 0xf6, |
399 | 0x0e, 0x8e, 0x4e, 0xce, 0x2e, 0xae, 0x6e, 0xee, |
400 | 0x1e, 0x9e, 0x5e, 0xde, 0x3e, 0xbe, 0x7e, 0xfe, |
401 | 0x01, 0x81, 0x41, 0xc1, 0x21, 0xa1, 0x61, 0xe1, |
402 | 0x11, 0x91, 0x51, 0xd1, 0x31, 0xb1, 0x71, 0xf1, |
403 | 0x09, 0x89, 0x49, 0xc9, 0x29, 0xa9, 0x69, 0xe9, |
404 | 0x19, 0x99, 0x59, 0xd9, 0x39, 0xb9, 0x79, 0xf9, |
405 | 0x05, 0x85, 0x45, 0xc5, 0x25, 0xa5, 0x65, 0xe5, |
406 | 0x15, 0x95, 0x55, 0xd5, 0x35, 0xb5, 0x75, 0xf5, |
407 | 0x0d, 0x8d, 0x4d, 0xcd, 0x2d, 0xad, 0x6d, 0xed, |
408 | 0x1d, 0x9d, 0x5d, 0xdd, 0x3d, 0xbd, 0x7d, 0xfd, |
409 | 0x03, 0x83, 0x43, 0xc3, 0x23, 0xa3, 0x63, 0xe3, |
410 | 0x13, 0x93, 0x53, 0xd3, 0x33, 0xb3, 0x73, 0xf3, |
411 | 0x0b, 0x8b, 0x4b, 0xcb, 0x2b, 0xab, 0x6b, 0xeb, |
412 | 0x1b, 0x9b, 0x5b, 0xdb, 0x3b, 0xbb, 0x7b, 0xfb, |
413 | 0x07, 0x87, 0x47, 0xc7, 0x27, 0xa7, 0x67, 0xe7, |
414 | 0x17, 0x97, 0x57, 0xd7, 0x37, 0xb7, 0x77, 0xf7, |
415 | 0x0f, 0x8f, 0x4f, 0xcf, 0x2f, 0xaf, 0x6f, 0xef, |
416 | 0x1f, 0x9f, 0x5f, 0xdf, 0x3f, 0xbf, 0x7f, 0xff, |
417 | }; |
418 | |
419 | typedef struct { |
d2371c81 |
420 | short code, extrabits; |
421 | int min, max; |
4ba9b64b |
422 | } coderecord; |
423 | |
424 | static const coderecord lencodes[] = { |
32874aea |
425 | {257, 0, 3, 3}, |
426 | {258, 0, 4, 4}, |
427 | {259, 0, 5, 5}, |
428 | {260, 0, 6, 6}, |
429 | {261, 0, 7, 7}, |
430 | {262, 0, 8, 8}, |
431 | {263, 0, 9, 9}, |
432 | {264, 0, 10, 10}, |
433 | {265, 1, 11, 12}, |
434 | {266, 1, 13, 14}, |
435 | {267, 1, 15, 16}, |
436 | {268, 1, 17, 18}, |
437 | {269, 2, 19, 22}, |
438 | {270, 2, 23, 26}, |
439 | {271, 2, 27, 30}, |
440 | {272, 2, 31, 34}, |
441 | {273, 3, 35, 42}, |
442 | {274, 3, 43, 50}, |
443 | {275, 3, 51, 58}, |
444 | {276, 3, 59, 66}, |
445 | {277, 4, 67, 82}, |
446 | {278, 4, 83, 98}, |
447 | {279, 4, 99, 114}, |
448 | {280, 4, 115, 130}, |
449 | {281, 5, 131, 162}, |
450 | {282, 5, 163, 194}, |
451 | {283, 5, 195, 226}, |
452 | {284, 5, 227, 257}, |
453 | {285, 0, 258, 258}, |
4ba9b64b |
454 | }; |
455 | |
456 | static const coderecord distcodes[] = { |
32874aea |
457 | {0, 0, 1, 1}, |
458 | {1, 0, 2, 2}, |
459 | {2, 0, 3, 3}, |
460 | {3, 0, 4, 4}, |
461 | {4, 1, 5, 6}, |
462 | {5, 1, 7, 8}, |
463 | {6, 2, 9, 12}, |
464 | {7, 2, 13, 16}, |
465 | {8, 3, 17, 24}, |
466 | {9, 3, 25, 32}, |
467 | {10, 4, 33, 48}, |
468 | {11, 4, 49, 64}, |
469 | {12, 5, 65, 96}, |
470 | {13, 5, 97, 128}, |
471 | {14, 6, 129, 192}, |
472 | {15, 6, 193, 256}, |
473 | {16, 7, 257, 384}, |
474 | {17, 7, 385, 512}, |
475 | {18, 8, 513, 768}, |
476 | {19, 8, 769, 1024}, |
477 | {20, 9, 1025, 1536}, |
478 | {21, 9, 1537, 2048}, |
479 | {22, 10, 2049, 3072}, |
480 | {23, 10, 3073, 4096}, |
481 | {24, 11, 4097, 6144}, |
482 | {25, 11, 6145, 8192}, |
483 | {26, 12, 8193, 12288}, |
484 | {27, 12, 12289, 16384}, |
485 | {28, 13, 16385, 24576}, |
486 | {29, 13, 24577, 32768}, |
4ba9b64b |
487 | }; |
488 | |
32874aea |
489 | static void zlib_literal(struct LZ77Context *ectx, unsigned char c) |
490 | { |
491 | struct Outbuf *out = (struct Outbuf *) ectx->userdata; |
4ba9b64b |
492 | |
6e9e9520 |
493 | if (out->comp_disabled) { |
32874aea |
494 | /* |
495 | * We're in an uncompressed block, so just output the byte. |
496 | */ |
497 | outbits(out, c, 8); |
498 | return; |
6e9e9520 |
499 | } |
500 | |
4ba9b64b |
501 | if (c <= 143) { |
32874aea |
502 | /* 0 through 143 are 8 bits long starting at 00110000. */ |
503 | outbits(out, mirrorbytes[0x30 + c], 8); |
4ba9b64b |
504 | } else { |
32874aea |
505 | /* 144 through 255 are 9 bits long starting at 110010000. */ |
506 | outbits(out, 1 + 2 * mirrorbytes[0x90 - 144 + c], 9); |
4ba9b64b |
507 | } |
508 | } |
509 | |
32874aea |
510 | static void zlib_match(struct LZ77Context *ectx, int distance, int len) |
511 | { |
4ba9b64b |
512 | const coderecord *d, *l; |
513 | int i, j, k; |
32874aea |
514 | struct Outbuf *out = (struct Outbuf *) ectx->userdata; |
6e9e9520 |
515 | |
516 | assert(!out->comp_disabled); |
517 | |
4ba9b64b |
518 | while (len > 0) { |
32874aea |
519 | int thislen; |
520 | |
4ba9b64b |
521 | /* |
522 | * We can transmit matches of lengths 3 through 258 |
523 | * inclusive. So if len exceeds 258, we must transmit in |
524 | * several steps, with 258 or less in each step. |
525 | * |
526 | * Specifically: if len >= 261, we can transmit 258 and be |
527 | * sure of having at least 3 left for the next step. And if |
528 | * len <= 258, we can just transmit len. But if len == 259 |
529 | * or 260, we must transmit len-3. |
530 | */ |
32874aea |
531 | thislen = (len > 260 ? 258 : len <= 258 ? len : len - 3); |
532 | len -= thislen; |
533 | |
534 | /* |
535 | * Binary-search to find which length code we're |
536 | * transmitting. |
537 | */ |
538 | i = -1; |
539 | j = sizeof(lencodes) / sizeof(*lencodes); |
2d466ffd |
540 | while (1) { |
541 | assert(j - i >= 2); |
32874aea |
542 | k = (j + i) / 2; |
543 | if (thislen < lencodes[k].min) |
544 | j = k; |
545 | else if (thislen > lencodes[k].max) |
546 | i = k; |
547 | else { |
548 | l = &lencodes[k]; |
549 | break; /* found it! */ |
550 | } |
551 | } |
552 | |
553 | /* |
554 | * Transmit the length code. 256-279 are seven bits |
555 | * starting at 0000000; 280-287 are eight bits starting at |
556 | * 11000000. |
557 | */ |
558 | if (l->code <= 279) { |
559 | outbits(out, mirrorbytes[(l->code - 256) * 2], 7); |
560 | } else { |
561 | outbits(out, mirrorbytes[0xc0 - 280 + l->code], 8); |
562 | } |
563 | |
564 | /* |
565 | * Transmit the extra bits. |
566 | */ |
567 | if (l->extrabits) |
568 | outbits(out, thislen - l->min, l->extrabits); |
569 | |
570 | /* |
571 | * Binary-search to find which distance code we're |
572 | * transmitting. |
573 | */ |
574 | i = -1; |
575 | j = sizeof(distcodes) / sizeof(*distcodes); |
2d466ffd |
576 | while (1) { |
577 | assert(j - i >= 2); |
32874aea |
578 | k = (j + i) / 2; |
579 | if (distance < distcodes[k].min) |
580 | j = k; |
581 | else if (distance > distcodes[k].max) |
582 | i = k; |
583 | else { |
584 | d = &distcodes[k]; |
585 | break; /* found it! */ |
586 | } |
587 | } |
588 | |
589 | /* |
590 | * Transmit the distance code. Five bits starting at 00000. |
591 | */ |
592 | outbits(out, mirrorbytes[d->code * 8], 5); |
593 | |
594 | /* |
595 | * Transmit the extra bits. |
596 | */ |
597 | if (d->extrabits) |
598 | outbits(out, distance - d->min, d->extrabits); |
4ba9b64b |
599 | } |
600 | } |
601 | |
5366aed8 |
602 | void *zlib_compress_init(void) |
32874aea |
603 | { |
4ba9b64b |
604 | struct Outbuf *out; |
3d88e64d |
605 | struct LZ77Context *ectx = snew(struct LZ77Context); |
4ba9b64b |
606 | |
5366aed8 |
607 | lz77_init(ectx); |
608 | ectx->literal = zlib_literal; |
609 | ectx->match = zlib_match; |
4ba9b64b |
610 | |
3d88e64d |
611 | out = snew(struct Outbuf); |
4ba9b64b |
612 | out->outbits = out->noutbits = 0; |
613 | out->firstblock = 1; |
6e9e9520 |
614 | out->comp_disabled = FALSE; |
5366aed8 |
615 | ectx->userdata = out; |
616 | |
617 | return ectx; |
618 | } |
4ba9b64b |
619 | |
5366aed8 |
620 | void zlib_compress_cleanup(void *handle) |
621 | { |
622 | struct LZ77Context *ectx = (struct LZ77Context *)handle; |
623 | sfree(ectx->userdata); |
29b1d0b3 |
624 | sfree(ectx->ictx); |
625 | sfree(ectx); |
4ba9b64b |
626 | } |
627 | |
6e9e9520 |
628 | /* |
629 | * Turn off actual LZ77 analysis for one block, to facilitate |
630 | * construction of a precise-length IGNORE packet. Returns the |
631 | * length adjustment (which is only valid for packets < 65536 |
632 | * bytes, but that seems reasonable enough). |
633 | */ |
6958ec09 |
634 | static int zlib_disable_compression(void *handle) |
32874aea |
635 | { |
5366aed8 |
636 | struct LZ77Context *ectx = (struct LZ77Context *)handle; |
637 | struct Outbuf *out = (struct Outbuf *) ectx->userdata; |
dae2b91f |
638 | int n; |
6e9e9520 |
639 | |
640 | out->comp_disabled = TRUE; |
641 | |
642 | n = 0; |
643 | /* |
644 | * If this is the first block, we will start by outputting two |
645 | * header bytes, and then three bits to begin an uncompressed |
646 | * block. This will cost three bytes (because we will start on |
647 | * a byte boundary, this is certain). |
648 | */ |
649 | if (out->firstblock) { |
32874aea |
650 | n = 3; |
6e9e9520 |
651 | } else { |
32874aea |
652 | /* |
653 | * Otherwise, we will output seven bits to close the |
654 | * previous static block, and _then_ three bits to begin an |
655 | * uncompressed block, and then flush the current byte. |
656 | * This may cost two bytes or three, depending on noutbits. |
657 | */ |
658 | n += (out->noutbits + 10) / 8; |
6e9e9520 |
659 | } |
660 | |
661 | /* |
662 | * Now we output four bytes for the length / ~length pair in |
663 | * the uncompressed block. |
664 | */ |
665 | n += 4; |
666 | |
667 | return n; |
668 | } |
669 | |
5366aed8 |
670 | int zlib_compress_block(void *handle, unsigned char *block, int len, |
32874aea |
671 | unsigned char **outblock, int *outlen) |
672 | { |
5366aed8 |
673 | struct LZ77Context *ectx = (struct LZ77Context *)handle; |
674 | struct Outbuf *out = (struct Outbuf *) ectx->userdata; |
6e9e9520 |
675 | int in_block; |
4ba9b64b |
676 | |
677 | out->outbuf = NULL; |
678 | out->outlen = out->outsize = 0; |
679 | |
680 | /* |
681 | * If this is the first block, output the Zlib (RFC1950) header |
682 | * bytes 78 9C. (Deflate compression, 32K window size, default |
683 | * algorithm.) |
684 | */ |
685 | if (out->firstblock) { |
32874aea |
686 | outbits(out, 0x9C78, 16); |
687 | out->firstblock = 0; |
6e9e9520 |
688 | |
32874aea |
689 | in_block = FALSE; |
2d466ffd |
690 | } else |
691 | in_block = TRUE; |
4ba9b64b |
692 | |
6e9e9520 |
693 | if (out->comp_disabled) { |
32874aea |
694 | if (in_block) |
695 | outbits(out, 0, 7); /* close static block */ |
696 | |
697 | while (len > 0) { |
698 | int blen = (len < 65535 ? len : 65535); |
699 | |
700 | /* |
701 | * Start a Deflate (RFC1951) uncompressed block. We |
e9122dbe |
702 | * transmit a zero bit (BFINAL=0), followed by two more |
703 | * zero bits (BTYPE=00). Of course these are in the |
704 | * wrong order (00 0), not that it matters. |
32874aea |
705 | */ |
706 | outbits(out, 0, 3); |
707 | |
708 | /* |
709 | * Output zero bits to align to a byte boundary. |
710 | */ |
711 | if (out->noutbits) |
712 | outbits(out, 0, 8 - out->noutbits); |
713 | |
714 | /* |
715 | * Output the block length, and then its one's |
716 | * complement. They're little-endian, so all we need to |
717 | * do is pass them straight to outbits() with bit count |
718 | * 16. |
719 | */ |
720 | outbits(out, blen, 16); |
721 | outbits(out, blen ^ 0xFFFF, 16); |
722 | |
723 | /* |
724 | * Do the `compression': we need to pass the data to |
725 | * lz77_compress so that it will be taken into account |
726 | * for subsequent (distance,length) pairs. But |
727 | * lz77_compress is passed FALSE, which means it won't |
728 | * actually find (or even look for) any matches; so |
729 | * every character will be passed straight to |
730 | * zlib_literal which will spot out->comp_disabled and |
731 | * emit in the uncompressed format. |
732 | */ |
5366aed8 |
733 | lz77_compress(ectx, block, blen, FALSE); |
32874aea |
734 | |
735 | len -= blen; |
736 | block += blen; |
737 | } |
738 | outbits(out, 2, 3); /* open new block */ |
6e9e9520 |
739 | } else { |
32874aea |
740 | if (!in_block) { |
741 | /* |
742 | * Start a Deflate (RFC1951) fixed-trees block. We |
743 | * transmit a zero bit (BFINAL=0), followed by a zero |
744 | * bit and a one bit (BTYPE=01). Of course these are in |
745 | * the wrong order (01 0). |
746 | */ |
747 | outbits(out, 2, 3); |
748 | } |
749 | |
750 | /* |
751 | * Do the compression. |
752 | */ |
5366aed8 |
753 | lz77_compress(ectx, block, len, TRUE); |
32874aea |
754 | |
755 | /* |
756 | * End the block (by transmitting code 256, which is |
757 | * 0000000 in fixed-tree mode), and transmit some empty |
758 | * blocks to ensure we have emitted the byte containing the |
759 | * last piece of genuine data. There are three ways we can |
760 | * do this: |
761 | * |
762 | * - Minimal flush. Output end-of-block and then open a |
763 | * new static block. This takes 9 bits, which is |
764 | * guaranteed to flush out the last genuine code in the |
765 | * closed block; but allegedly zlib can't handle it. |
766 | * |
767 | * - Zlib partial flush. Output EOB, open and close an |
768 | * empty static block, and _then_ open the new block. |
769 | * This is the best zlib can handle. |
770 | * |
771 | * - Zlib sync flush. Output EOB, then an empty |
772 | * _uncompressed_ block (000, then sync to byte |
773 | * boundary, then send bytes 00 00 FF FF). Then open the |
774 | * new block. |
775 | * |
776 | * For the moment, we will use Zlib partial flush. |
777 | */ |
778 | outbits(out, 0, 7); /* close block */ |
779 | outbits(out, 2, 3 + 7); /* empty static block */ |
780 | outbits(out, 2, 3); /* open new block */ |
6e9e9520 |
781 | } |
782 | |
783 | out->comp_disabled = FALSE; |
4ba9b64b |
784 | |
785 | *outblock = out->outbuf; |
786 | *outlen = out->outlen; |
787 | |
788 | return 1; |
789 | } |
790 | |
791 | /* ---------------------------------------------------------------------- |
792 | * Zlib decompression. Of course, even though our compressor always |
793 | * uses static trees, our _decompressor_ has to be capable of |
794 | * handling dynamic trees if it sees them. |
795 | */ |
796 | |
797 | /* |
798 | * The way we work the Huffman decode is to have a table lookup on |
799 | * the first N bits of the input stream (in the order they arrive, |
800 | * of course, i.e. the first bit of the Huffman code is in bit 0). |
801 | * Each table entry lists the number of bits to consume, plus |
802 | * either an output code or a pointer to a secondary table. |
803 | */ |
804 | struct zlib_table; |
805 | struct zlib_tableentry; |
806 | |
807 | struct zlib_tableentry { |
808 | unsigned char nbits; |
d2371c81 |
809 | short code; |
4ba9b64b |
810 | struct zlib_table *nexttable; |
811 | }; |
812 | |
813 | struct zlib_table { |
32874aea |
814 | int mask; /* mask applied to input bit stream */ |
4ba9b64b |
815 | struct zlib_tableentry *table; |
816 | }; |
817 | |
818 | #define MAXCODELEN 16 |
819 | #define MAXSYMS 288 |
820 | |
821 | /* |
822 | * Build a single-level decode table for elements |
823 | * [minlength,maxlength) of the provided code/length tables, and |
824 | * recurse to build subtables. |
825 | */ |
826 | static struct zlib_table *zlib_mkonetab(int *codes, unsigned char *lengths, |
32874aea |
827 | int nsyms, |
828 | int pfx, int pfxbits, int bits) |
829 | { |
3d88e64d |
830 | struct zlib_table *tab = snew(struct zlib_table); |
4ba9b64b |
831 | int pfxmask = (1 << pfxbits) - 1; |
832 | int nbits, i, j, code; |
833 | |
3d88e64d |
834 | tab->table = snewn(1 << bits, struct zlib_tableentry); |
4ba9b64b |
835 | tab->mask = (1 << bits) - 1; |
836 | |
837 | for (code = 0; code <= tab->mask; code++) { |
32874aea |
838 | tab->table[code].code = -1; |
839 | tab->table[code].nbits = 0; |
840 | tab->table[code].nexttable = NULL; |
4ba9b64b |
841 | } |
842 | |
843 | for (i = 0; i < nsyms; i++) { |
32874aea |
844 | if (lengths[i] <= pfxbits || (codes[i] & pfxmask) != pfx) |
845 | continue; |
846 | code = (codes[i] >> pfxbits) & tab->mask; |
847 | for (j = code; j <= tab->mask; j += 1 << (lengths[i] - pfxbits)) { |
848 | tab->table[j].code = i; |
849 | nbits = lengths[i] - pfxbits; |
850 | if (tab->table[j].nbits < nbits) |
851 | tab->table[j].nbits = nbits; |
852 | } |
4ba9b64b |
853 | } |
854 | for (code = 0; code <= tab->mask; code++) { |
32874aea |
855 | if (tab->table[code].nbits <= bits) |
856 | continue; |
857 | /* Generate a subtable. */ |
858 | tab->table[code].code = -1; |
859 | nbits = tab->table[code].nbits - bits; |
860 | if (nbits > 7) |
861 | nbits = 7; |
862 | tab->table[code].nbits = bits; |
863 | tab->table[code].nexttable = zlib_mkonetab(codes, lengths, nsyms, |
864 | pfx | (code << pfxbits), |
865 | pfxbits + bits, nbits); |
4ba9b64b |
866 | } |
867 | |
868 | return tab; |
869 | } |
870 | |
871 | /* |
872 | * Build a decode table, given a set of Huffman tree lengths. |
873 | */ |
32874aea |
874 | static struct zlib_table *zlib_mktable(unsigned char *lengths, |
875 | int nlengths) |
876 | { |
4ba9b64b |
877 | int count[MAXCODELEN], startcode[MAXCODELEN], codes[MAXSYMS]; |
878 | int code, maxlen; |
879 | int i, j; |
880 | |
881 | /* Count the codes of each length. */ |
882 | maxlen = 0; |
32874aea |
883 | for (i = 1; i < MAXCODELEN; i++) |
884 | count[i] = 0; |
4ba9b64b |
885 | for (i = 0; i < nlengths; i++) { |
32874aea |
886 | count[lengths[i]]++; |
887 | if (maxlen < lengths[i]) |
888 | maxlen = lengths[i]; |
4ba9b64b |
889 | } |
890 | /* Determine the starting code for each length block. */ |
891 | code = 0; |
892 | for (i = 1; i < MAXCODELEN; i++) { |
32874aea |
893 | startcode[i] = code; |
894 | code += count[i]; |
895 | code <<= 1; |
4ba9b64b |
896 | } |
897 | /* Determine the code for each symbol. Mirrored, of course. */ |
898 | for (i = 0; i < nlengths; i++) { |
32874aea |
899 | code = startcode[lengths[i]]++; |
900 | codes[i] = 0; |
901 | for (j = 0; j < lengths[i]; j++) { |
902 | codes[i] = (codes[i] << 1) | (code & 1); |
903 | code >>= 1; |
904 | } |
4ba9b64b |
905 | } |
906 | |
907 | /* |
908 | * Now we have the complete list of Huffman codes. Build a |
909 | * table. |
910 | */ |
911 | return zlib_mkonetab(codes, lengths, nlengths, 0, 0, |
32874aea |
912 | maxlen < 9 ? maxlen : 9); |
4ba9b64b |
913 | } |
914 | |
32874aea |
915 | static int zlib_freetable(struct zlib_table **ztab) |
916 | { |
dda46d54 |
917 | struct zlib_table *tab; |
918 | int code; |
919 | |
920 | if (ztab == NULL) |
921 | return -1; |
922 | |
923 | if (*ztab == NULL) |
924 | return 0; |
925 | |
926 | tab = *ztab; |
927 | |
928 | for (code = 0; code <= tab->mask; code++) |
929 | if (tab->table[code].nexttable != NULL) |
930 | zlib_freetable(&tab->table[code].nexttable); |
931 | |
932 | sfree(tab->table); |
933 | tab->table = NULL; |
934 | |
935 | sfree(tab); |
936 | *ztab = NULL; |
937 | |
32874aea |
938 | return (0); |
dda46d54 |
939 | } |
940 | |
5366aed8 |
941 | struct zlib_decompress_ctx { |
4ba9b64b |
942 | struct zlib_table *staticlentable, *staticdisttable; |
943 | struct zlib_table *currlentable, *currdisttable, *lenlentable; |
944 | enum { |
32874aea |
945 | START, OUTSIDEBLK, |
946 | TREES_HDR, TREES_LENLEN, TREES_LEN, TREES_LENREP, |
947 | INBLK, GOTLENSYM, GOTLEN, GOTDISTSYM, |
948 | UNCOMP_LEN, UNCOMP_NLEN, UNCOMP_DATA |
4ba9b64b |
949 | } state; |
32874aea |
950 | int sym, hlit, hdist, hclen, lenptr, lenextrabits, lenaddon, len, |
951 | lenrep; |
4ba9b64b |
952 | int uncomplen; |
953 | unsigned char lenlen[19]; |
32874aea |
954 | unsigned char lengths[286 + 32]; |
4ba9b64b |
955 | unsigned long bits; |
956 | int nbits; |
957 | unsigned char window[WINSIZE]; |
958 | int winpos; |
959 | unsigned char *outblk; |
960 | int outlen, outsize; |
5366aed8 |
961 | }; |
4ba9b64b |
962 | |
5366aed8 |
963 | void *zlib_decompress_init(void) |
32874aea |
964 | { |
3d88e64d |
965 | struct zlib_decompress_ctx *dctx = snew(struct zlib_decompress_ctx); |
4ba9b64b |
966 | unsigned char lengths[288]; |
5366aed8 |
967 | |
4ba9b64b |
968 | memset(lengths, 8, 144); |
32874aea |
969 | memset(lengths + 144, 9, 256 - 144); |
970 | memset(lengths + 256, 7, 280 - 256); |
971 | memset(lengths + 280, 8, 288 - 280); |
5366aed8 |
972 | dctx->staticlentable = zlib_mktable(lengths, 288); |
4ba9b64b |
973 | memset(lengths, 5, 32); |
5366aed8 |
974 | dctx->staticdisttable = zlib_mktable(lengths, 32); |
975 | dctx->state = START; /* even before header */ |
976 | dctx->currlentable = dctx->currdisttable = dctx->lenlentable = NULL; |
977 | dctx->bits = 0; |
978 | dctx->nbits = 0; |
a8327734 |
979 | dctx->winpos = 0; |
5366aed8 |
980 | |
981 | return dctx; |
982 | } |
983 | |
984 | void zlib_decompress_cleanup(void *handle) |
985 | { |
986 | struct zlib_decompress_ctx *dctx = (struct zlib_decompress_ctx *)handle; |
29b1d0b3 |
987 | |
5366aed8 |
988 | if (dctx->currlentable && dctx->currlentable != dctx->staticlentable) |
989 | zlib_freetable(&dctx->currlentable); |
990 | if (dctx->currdisttable && dctx->currdisttable != dctx->staticdisttable) |
991 | zlib_freetable(&dctx->currdisttable); |
992 | if (dctx->lenlentable) |
993 | zlib_freetable(&dctx->lenlentable); |
29b1d0b3 |
994 | zlib_freetable(&dctx->staticlentable); |
995 | zlib_freetable(&dctx->staticdisttable); |
5366aed8 |
996 | sfree(dctx); |
4ba9b64b |
997 | } |
998 | |
6958ec09 |
999 | static int zlib_huflookup(unsigned long *bitsp, int *nbitsp, |
32874aea |
1000 | struct zlib_table *tab) |
1001 | { |
4ba9b64b |
1002 | unsigned long bits = *bitsp; |
1003 | int nbits = *nbitsp; |
1004 | while (1) { |
32874aea |
1005 | struct zlib_tableentry *ent; |
1006 | ent = &tab->table[bits & tab->mask]; |
1007 | if (ent->nbits > nbits) |
1008 | return -1; /* not enough data */ |
1009 | bits >>= ent->nbits; |
1010 | nbits -= ent->nbits; |
1011 | if (ent->code == -1) |
1012 | tab = ent->nexttable; |
1013 | else { |
1014 | *bitsp = bits; |
1015 | *nbitsp = nbits; |
1016 | return ent->code; |
1017 | } |
36b8d9bb |
1018 | |
1019 | if (!tab) { |
1020 | /* |
1021 | * There was a missing entry in the table, presumably |
1022 | * due to an invalid Huffman table description, and the |
1023 | * subsequent data has attempted to use the missing |
1024 | * entry. Return a decoding failure. |
1025 | */ |
1026 | return -2; |
1027 | } |
4ba9b64b |
1028 | } |
1029 | } |
1030 | |
5366aed8 |
1031 | static void zlib_emit_char(struct zlib_decompress_ctx *dctx, int c) |
32874aea |
1032 | { |
5366aed8 |
1033 | dctx->window[dctx->winpos] = c; |
1034 | dctx->winpos = (dctx->winpos + 1) & (WINSIZE - 1); |
1035 | if (dctx->outlen >= dctx->outsize) { |
1036 | dctx->outsize = dctx->outlen + 512; |
3d88e64d |
1037 | dctx->outblk = sresize(dctx->outblk, dctx->outsize, unsigned char); |
4ba9b64b |
1038 | } |
5366aed8 |
1039 | dctx->outblk[dctx->outlen++] = c; |
4ba9b64b |
1040 | } |
1041 | |
5366aed8 |
1042 | #define EATBITS(n) ( dctx->nbits -= (n), dctx->bits >>= (n) ) |
4ba9b64b |
1043 | |
5366aed8 |
1044 | int zlib_decompress_block(void *handle, unsigned char *block, int len, |
32874aea |
1045 | unsigned char **outblock, int *outlen) |
1046 | { |
5366aed8 |
1047 | struct zlib_decompress_ctx *dctx = (struct zlib_decompress_ctx *)handle; |
4ba9b64b |
1048 | const coderecord *rec; |
52912ff3 |
1049 | int code, blktype, rep, dist, nlen, header; |
4ba9b64b |
1050 | static const unsigned char lenlenmap[] = { |
32874aea |
1051 | 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 |
4ba9b64b |
1052 | }; |
1053 | |
52912ff3 |
1054 | dctx->outblk = snewn(256, unsigned char); |
1055 | dctx->outsize = 256; |
1056 | dctx->outlen = 0; |
4ba9b64b |
1057 | |
5366aed8 |
1058 | while (len > 0 || dctx->nbits > 0) { |
1059 | while (dctx->nbits < 24 && len > 0) { |
1060 | dctx->bits |= (*block++) << dctx->nbits; |
1061 | dctx->nbits += 8; |
32874aea |
1062 | len--; |
1063 | } |
5366aed8 |
1064 | switch (dctx->state) { |
32874aea |
1065 | case START: |
52912ff3 |
1066 | /* Expect 16-bit zlib header. */ |
5366aed8 |
1067 | if (dctx->nbits < 16) |
32874aea |
1068 | goto finished; /* done all we can */ |
52912ff3 |
1069 | |
1070 | /* |
1071 | * The header is stored as a big-endian 16-bit integer, |
1072 | * in contrast to the general little-endian policy in |
1073 | * the rest of the format :-( |
1074 | */ |
1075 | header = (((dctx->bits & 0xFF00) >> 8) | |
1076 | ((dctx->bits & 0x00FF) << 8)); |
1077 | EATBITS(16); |
1078 | |
1079 | /* |
1080 | * Check the header: |
1081 | * |
1082 | * - bits 8-11 should be 1000 (Deflate/RFC1951) |
1083 | * - bits 12-15 should be at most 0111 (window size) |
1084 | * - bit 5 should be zero (no dictionary present) |
1085 | * - we don't care about bits 6-7 (compression rate) |
1086 | * - bits 0-4 should be set up to make the whole thing |
1087 | * a multiple of 31 (checksum). |
1088 | */ |
1089 | if ((header & 0x0F00) != 0x0800 || |
1090 | (header & 0xF000) > 0x7000 || |
1091 | (header & 0x0020) != 0x0000 || |
1092 | (header % 31) != 0) |
1093 | goto decode_error; |
1094 | |
5366aed8 |
1095 | dctx->state = OUTSIDEBLK; |
32874aea |
1096 | break; |
1097 | case OUTSIDEBLK: |
1098 | /* Expect 3-bit block header. */ |
5366aed8 |
1099 | if (dctx->nbits < 3) |
32874aea |
1100 | goto finished; /* done all we can */ |
1101 | EATBITS(1); |
5366aed8 |
1102 | blktype = dctx->bits & 3; |
32874aea |
1103 | EATBITS(2); |
1104 | if (blktype == 0) { |
5366aed8 |
1105 | int to_eat = dctx->nbits & 7; |
1106 | dctx->state = UNCOMP_LEN; |
4ba9b64b |
1107 | EATBITS(to_eat); /* align to byte boundary */ |
32874aea |
1108 | } else if (blktype == 1) { |
5366aed8 |
1109 | dctx->currlentable = dctx->staticlentable; |
1110 | dctx->currdisttable = dctx->staticdisttable; |
1111 | dctx->state = INBLK; |
32874aea |
1112 | } else if (blktype == 2) { |
5366aed8 |
1113 | dctx->state = TREES_HDR; |
32874aea |
1114 | } |
1115 | break; |
1116 | case TREES_HDR: |
1117 | /* |
1118 | * Dynamic block header. Five bits of HLIT, five of |
1119 | * HDIST, four of HCLEN. |
1120 | */ |
5366aed8 |
1121 | if (dctx->nbits < 5 + 5 + 4) |
32874aea |
1122 | goto finished; /* done all we can */ |
5366aed8 |
1123 | dctx->hlit = 257 + (dctx->bits & 31); |
32874aea |
1124 | EATBITS(5); |
5366aed8 |
1125 | dctx->hdist = 1 + (dctx->bits & 31); |
32874aea |
1126 | EATBITS(5); |
5366aed8 |
1127 | dctx->hclen = 4 + (dctx->bits & 15); |
32874aea |
1128 | EATBITS(4); |
5366aed8 |
1129 | dctx->lenptr = 0; |
1130 | dctx->state = TREES_LENLEN; |
1131 | memset(dctx->lenlen, 0, sizeof(dctx->lenlen)); |
32874aea |
1132 | break; |
1133 | case TREES_LENLEN: |
5366aed8 |
1134 | if (dctx->nbits < 3) |
32874aea |
1135 | goto finished; |
5366aed8 |
1136 | while (dctx->lenptr < dctx->hclen && dctx->nbits >= 3) { |
1137 | dctx->lenlen[lenlenmap[dctx->lenptr++]] = |
1138 | (unsigned char) (dctx->bits & 7); |
32874aea |
1139 | EATBITS(3); |
1140 | } |
5366aed8 |
1141 | if (dctx->lenptr == dctx->hclen) { |
1142 | dctx->lenlentable = zlib_mktable(dctx->lenlen, 19); |
1143 | dctx->state = TREES_LEN; |
1144 | dctx->lenptr = 0; |
32874aea |
1145 | } |
1146 | break; |
1147 | case TREES_LEN: |
5366aed8 |
1148 | if (dctx->lenptr >= dctx->hlit + dctx->hdist) { |
1149 | dctx->currlentable = zlib_mktable(dctx->lengths, dctx->hlit); |
1150 | dctx->currdisttable = zlib_mktable(dctx->lengths + dctx->hlit, |
1151 | dctx->hdist); |
1152 | zlib_freetable(&dctx->lenlentable); |
1153 | dctx->lenlentable = NULL; |
1154 | dctx->state = INBLK; |
32874aea |
1155 | break; |
1156 | } |
1157 | code = |
5366aed8 |
1158 | zlib_huflookup(&dctx->bits, &dctx->nbits, dctx->lenlentable); |
32874aea |
1159 | if (code == -1) |
1160 | goto finished; |
36b8d9bb |
1161 | if (code == -2) |
1162 | goto decode_error; |
32874aea |
1163 | if (code < 16) |
5366aed8 |
1164 | dctx->lengths[dctx->lenptr++] = code; |
32874aea |
1165 | else { |
5366aed8 |
1166 | dctx->lenextrabits = (code == 16 ? 2 : code == 17 ? 3 : 7); |
1167 | dctx->lenaddon = (code == 18 ? 11 : 3); |
1168 | dctx->lenrep = (code == 16 && dctx->lenptr > 0 ? |
1169 | dctx->lengths[dctx->lenptr - 1] : 0); |
1170 | dctx->state = TREES_LENREP; |
32874aea |
1171 | } |
1172 | break; |
1173 | case TREES_LENREP: |
5366aed8 |
1174 | if (dctx->nbits < dctx->lenextrabits) |
32874aea |
1175 | goto finished; |
1176 | rep = |
5366aed8 |
1177 | dctx->lenaddon + |
1178 | (dctx->bits & ((1 << dctx->lenextrabits) - 1)); |
1179 | EATBITS(dctx->lenextrabits); |
1180 | while (rep > 0 && dctx->lenptr < dctx->hlit + dctx->hdist) { |
1181 | dctx->lengths[dctx->lenptr] = dctx->lenrep; |
1182 | dctx->lenptr++; |
32874aea |
1183 | rep--; |
1184 | } |
5366aed8 |
1185 | dctx->state = TREES_LEN; |
32874aea |
1186 | break; |
1187 | case INBLK: |
1188 | code = |
5366aed8 |
1189 | zlib_huflookup(&dctx->bits, &dctx->nbits, dctx->currlentable); |
32874aea |
1190 | if (code == -1) |
1191 | goto finished; |
36b8d9bb |
1192 | if (code == -2) |
1193 | goto decode_error; |
32874aea |
1194 | if (code < 256) |
5366aed8 |
1195 | zlib_emit_char(dctx, code); |
32874aea |
1196 | else if (code == 256) { |
5366aed8 |
1197 | dctx->state = OUTSIDEBLK; |
1198 | if (dctx->currlentable != dctx->staticlentable) { |
1199 | zlib_freetable(&dctx->currlentable); |
1200 | dctx->currlentable = NULL; |
1201 | } |
1202 | if (dctx->currdisttable != dctx->staticdisttable) { |
1203 | zlib_freetable(&dctx->currdisttable); |
1204 | dctx->currdisttable = NULL; |
1205 | } |
32874aea |
1206 | } else if (code < 286) { /* static tree can give >285; ignore */ |
5366aed8 |
1207 | dctx->state = GOTLENSYM; |
1208 | dctx->sym = code; |
32874aea |
1209 | } |
1210 | break; |
1211 | case GOTLENSYM: |
5366aed8 |
1212 | rec = &lencodes[dctx->sym - 257]; |
1213 | if (dctx->nbits < rec->extrabits) |
32874aea |
1214 | goto finished; |
5366aed8 |
1215 | dctx->len = |
1216 | rec->min + (dctx->bits & ((1 << rec->extrabits) - 1)); |
32874aea |
1217 | EATBITS(rec->extrabits); |
5366aed8 |
1218 | dctx->state = GOTLEN; |
32874aea |
1219 | break; |
1220 | case GOTLEN: |
1221 | code = |
5366aed8 |
1222 | zlib_huflookup(&dctx->bits, &dctx->nbits, |
1223 | dctx->currdisttable); |
32874aea |
1224 | if (code == -1) |
1225 | goto finished; |
36b8d9bb |
1226 | if (code == -2) |
1227 | goto decode_error; |
5366aed8 |
1228 | dctx->state = GOTDISTSYM; |
1229 | dctx->sym = code; |
32874aea |
1230 | break; |
1231 | case GOTDISTSYM: |
5366aed8 |
1232 | rec = &distcodes[dctx->sym]; |
1233 | if (dctx->nbits < rec->extrabits) |
32874aea |
1234 | goto finished; |
5366aed8 |
1235 | dist = rec->min + (dctx->bits & ((1 << rec->extrabits) - 1)); |
32874aea |
1236 | EATBITS(rec->extrabits); |
5366aed8 |
1237 | dctx->state = INBLK; |
1238 | while (dctx->len--) |
1239 | zlib_emit_char(dctx, dctx->window[(dctx->winpos - dist) & |
1240 | (WINSIZE - 1)]); |
4ba9b64b |
1241 | break; |
1242 | case UNCOMP_LEN: |
1243 | /* |
1244 | * Uncompressed block. We expect to see a 16-bit LEN. |
1245 | */ |
5366aed8 |
1246 | if (dctx->nbits < 16) |
4ba9b64b |
1247 | goto finished; |
5366aed8 |
1248 | dctx->uncomplen = dctx->bits & 0xFFFF; |
4ba9b64b |
1249 | EATBITS(16); |
5366aed8 |
1250 | dctx->state = UNCOMP_NLEN; |
4ba9b64b |
1251 | break; |
1252 | case UNCOMP_NLEN: |
1253 | /* |
1254 | * Uncompressed block. We expect to see a 16-bit NLEN, |
1255 | * which should be the one's complement of the previous |
1256 | * LEN. |
1257 | */ |
5366aed8 |
1258 | if (dctx->nbits < 16) |
4ba9b64b |
1259 | goto finished; |
5366aed8 |
1260 | nlen = dctx->bits & 0xFFFF; |
4ba9b64b |
1261 | EATBITS(16); |
5366aed8 |
1262 | if (dctx->uncomplen == 0) |
1263 | dctx->state = OUTSIDEBLK; /* block is empty */ |
7b4fa1b4 |
1264 | else |
5366aed8 |
1265 | dctx->state = UNCOMP_DATA; |
4ba9b64b |
1266 | break; |
1267 | case UNCOMP_DATA: |
5366aed8 |
1268 | if (dctx->nbits < 8) |
4ba9b64b |
1269 | goto finished; |
5366aed8 |
1270 | zlib_emit_char(dctx, dctx->bits & 0xFF); |
4ba9b64b |
1271 | EATBITS(8); |
5366aed8 |
1272 | if (--dctx->uncomplen == 0) |
1273 | dctx->state = OUTSIDEBLK; /* end of uncompressed block */ |
4ba9b64b |
1274 | break; |
32874aea |
1275 | } |
4ba9b64b |
1276 | } |
1277 | |
32874aea |
1278 | finished: |
5366aed8 |
1279 | *outblock = dctx->outblk; |
1280 | *outlen = dctx->outlen; |
4ba9b64b |
1281 | return 1; |
36b8d9bb |
1282 | |
1283 | decode_error: |
1284 | sfree(dctx->outblk); |
1285 | *outblock = dctx->outblk = NULL; |
1286 | *outlen = 0; |
1287 | return 0; |
4ba9b64b |
1288 | } |
1289 | |
52912ff3 |
1290 | #ifdef ZLIB_STANDALONE |
1291 | |
1292 | #include <stdio.h> |
1293 | #include <string.h> |
1294 | |
1295 | int main(int argc, char **argv) |
1296 | { |
1297 | unsigned char buf[16], *outbuf; |
1298 | int ret, outlen; |
1299 | void *handle; |
1300 | int noheader = FALSE, opts = TRUE; |
1301 | char *filename = NULL; |
1302 | FILE *fp; |
1303 | |
1304 | while (--argc) { |
1305 | char *p = *++argv; |
1306 | |
1307 | if (p[0] == '-' && opts) { |
1308 | if (!strcmp(p, "-d")) |
1309 | noheader = TRUE; |
1310 | else if (!strcmp(p, "--")) |
1311 | opts = FALSE; /* next thing is filename */ |
1312 | else { |
1313 | fprintf(stderr, "unknown command line option '%s'\n", p); |
1314 | return 1; |
1315 | } |
1316 | } else if (!filename) { |
1317 | filename = p; |
1318 | } else { |
1319 | fprintf(stderr, "can only handle one filename\n"); |
1320 | return 1; |
1321 | } |
1322 | } |
1323 | |
1324 | handle = zlib_decompress_init(); |
1325 | |
1326 | if (noheader) { |
1327 | /* |
1328 | * Provide missing zlib header if -d was specified. |
1329 | */ |
1330 | zlib_decompress_block(handle, "\x78\x9C", 2, &outbuf, &outlen); |
1331 | assert(outlen == 0); |
1332 | } |
1333 | |
1334 | if (filename) |
1335 | fp = fopen(filename, "rb"); |
1336 | else |
1337 | fp = stdin; |
1338 | |
1339 | if (!fp) { |
1340 | assert(filename); |
1341 | fprintf(stderr, "unable to open '%s'\n", filename); |
1342 | return 1; |
1343 | } |
1344 | |
1345 | while (1) { |
1346 | ret = fread(buf, 1, sizeof(buf), fp); |
1347 | if (ret <= 0) |
1348 | break; |
1349 | zlib_decompress_block(handle, buf, ret, &outbuf, &outlen); |
1350 | if (outbuf) { |
1351 | if (outlen) |
1352 | fwrite(outbuf, 1, outlen, stdout); |
1353 | sfree(outbuf); |
1354 | } else { |
1355 | fprintf(stderr, "decoding error\n"); |
1356 | return 1; |
1357 | } |
1358 | } |
1359 | |
1360 | zlib_decompress_cleanup(handle); |
1361 | |
1362 | if (filename) |
1363 | fclose(fp); |
1364 | |
1365 | return 0; |
1366 | } |
1367 | |
1368 | #else |
1369 | |
4ba9b64b |
1370 | const struct ssh_compress ssh_zlib = { |
1371 | "zlib", |
1372 | zlib_compress_init, |
5366aed8 |
1373 | zlib_compress_cleanup, |
4ba9b64b |
1374 | zlib_compress_block, |
1375 | zlib_decompress_init, |
5366aed8 |
1376 | zlib_decompress_cleanup, |
6e9e9520 |
1377 | zlib_decompress_block, |
5366aed8 |
1378 | zlib_disable_compression, |
1379 | "zlib (RFC1950)" |
4ba9b64b |
1380 | }; |
52912ff3 |
1381 | |
1382 | #endif |